On Sunday, 29 December 2019 at 14:41:46 UTC, Russel Winder wrote:
On Sat, 2019-12-28 at 22:01 +0000, p.shkadzko via
Digitalmars-d-learn
wrote:
[…]
p.s. I found it quite satisfying that D does not really need
an IDE, you will be fine even with nano.
The fundamental issue with these all battery included fancy IDE's
(especially in Java) is that they tend to become dependencies of
the projects themselves.
How many times have I seen in my professionnal world, projects
that required specific versions of Eclipse with specific versions
of extensions and libraries?
At my work we have exactly currently the problem. One developer
wrote one of the desktop apps and now left the company. My
colleagues of that department are now struggling to maintain the
app as it used some specific GUI libs linked to some Eclipse
version and they are nowhere to be found. You may object that
it's a problem of the project management and I would agree. It
was the management error to let the developer choose the IDE
solution in the first place. A more classical/portable approach
would have been preferable.
Furthermore, it is extremely annoying that these IDE change over
time and all the fancy stuff gets stale and changed with other
stuff that gets stale after time.
Visual Studio is one of the worst offenders in that category.
Every 5 years it changes so much that everything learnt before
can be thrown away.
IDE's work well for scenarios that the developers of the IDE
thought of. Anything a little bit different requires changes that
are either impossible to model or require intimate knowledge of
the functionning of the IDE. Visual Studio comes to mind again of
an example where that is horribly painful (I do not even mention
the difficulty to even install such behemoth programs on our
corporate laptops which are behind stupid proxies and follow
annoying corporate policy rules).
